NEIGHBORHOOD LOCATION EFFICIENCY AND HOUSEHOLD TRANSPORTATION COSTS The 15 th TRB National Transportation Planning Applications Conference Innovative Tools in Transportation and Land Use Planning May 18, 2015 – Atlantic City, NJ Peter Haas, Ph.D. – Chief Research Scientist, Center for Neighborhood Technology

ABOUT CNT  CNT is a national hub for research, strategies and solutions to help cities use resources more efficiently and equitably.  We believe solving problems like poverty, climate change and urban sprawl starts with making neighborhoods, cities and regions work better.

TRANSPORTATION IS A HOUSEHOLD’S SECOND-LARGEST EXPENSE AFTER HOUSING Photo Credit: straightedge217/Flickr, Creative Commons License

TRANSPORTATION COSTS DEPEND ON PLACE Photo Credit: Ann Fisher/Flickr, Creative Commons License

Household Transportation Cost Model Households know their housing costs; An accurate estimation of household transportation costs related to a household location is elusive; Therefore CNT created a household transportation cost model. It is: An aggregate place based model. Based on US Census Block Groups as its unit of analysis. Estimating the cost of all driving, and transit. Predictive model to be used for the H+T Index. National in scope for all metropolitan and micropolitan areas. It is not: A travel demand model. Derived from travel survey data. A journey to work model. Necessarily a parsimonious model. A rural cost model.

Household Transportation Cost Model

Dependent Variables Auto Ownership 2013 ACS Average vehicles per occupied housing unit (HH) Transit Use No ubiquitous measure of transit use is available at the block group level, As proxy, the percent of commuters utilizing public transit as means of transportation to work.

Dependent Variables (Continued) Auto Use Odometer readings from 2010 through 2012 odometer readings were acquired in Illinois for the Chicago and St. Louis metro areas. Data were matched for over 660,000 records (two records for each individual vehicle identification number (VIN)) and the change provided VMT estimates. The dataset represents a diverse set of place types from rural areas to large cities, and provides a very good data set to calibrate the model. Data obtained were geographically identified with ZIP+4 TM and then assigned to Census block groups. The final value of VMT includes an additional factor of eight percent to compensate for the fact that the vehicles in this sample were all five years old or older.

Independent Variables – Household V ARIABLE D ESCRIPTION D ATA S OURCE T YPE M EDIAN HH I NCOME M EDIAN HOUSEHOLD I NCOME IN THE BLOCK GROUP 2013 ACSH OUSEHOLD C OMMUTERS /HH W ORKERS PER HOUSEHOLD WHO DO NOT WORK AT HOME 2013 ACSH OUSEHOLD A VG. HH S IZE A VERAGE NUMBER OF PEOPLE PER HOUSEHOLD 2013 ACSH OUSEHOLD

Independent Variables – Neighborhood V ARIABLE D ESCRIPTION D ATA S OURCE T YPE G ROSS H OUSEHOLD D ENSITY N UMBER OF HOUSEHOLDS DIVIDED BY THE LAND AREA IN THE C ENSUS BLOCK GROUP 2013 ACS, TIGER/L INE FILES N EIGHBORHOOD C HARACTERISTIC ( HOUSING DENSITY ) R EGIONAL H OUSEHOLD I NTENSITY H OUSEHOLDS SUMMED DIVIDED BY THE DISTANCE SQUARED IN MILES BETWEEN BLOCK GROUP BY ( THE HOUSEHOLDS IN THE BLOCK GROUP ARE NOT INCLUDED ) 2013 ACS, TIGER/L INE FILES N EIGHBORHOOD C HARACTERISTIC ( HOUSING DENSITY ) F RACTION OF S INGLE F AMILY D ETACHED H OUSING F RACTION OF SINGLE FAMILY DETACHED HOUSEHOLDS IN THE BLOCK GROUP 2013 ACS, TIGER/L INE FILES N EIGHBORHOOD C HARACTERISTIC ( HOUSING DENSITY ) E MPLOYMENT A CCESS I NDEX J OBS SUMMED BY BLOCKS DIVIDED BY THE DISTANCE SQUARED IN MILES ( IF LESS THAN ONE MILE NOT SCALED ) C ENSUS LEHD-LODES N EIGHBORHOOD C HARACTERISTIC ( EMPLOYMENT ) E MPLOYMENT M IX I NDEX N UMBER OF BLOCK PER ACRE W EIGHTED SUM OF 13 DIFFERENT EMPLOYMENT TYPES EACH SCALED BY A COEFFICIENT THAT ARE OPTIMIZED USING TRANSIT USE C ENSUS LEHD-LODES N EIGHBORHOOD C HARACTERISTIC ( EMPLOYMENT ) B LOCK D ENSITY N UMBER OF BLOCK PER ACRE TIGER/L INE FILES N EIGHBORHOOD C HARACTERISTIC ( WALKABILITY ) T RANSIT C ONNECTIVITY I NDEX S UM OF BUSES / TRAINS PER WEEK SCALED BY OVERLAP OF 1/8 MILE RINGS ABOUT EVERY STOP THAT INTERSECTS THE BLOCK GROUP CNT A LL T RANSIT N EIGHBORHOOD C HARACTERISTIC ( TRANSIT ) A VERAGE A VAILABLE T RANSIT T RIPS PER W EEK N UMBER OF POSSIBLE TRANSIT RIDES WITHIN THE BLOCK GROUP AND A ¼ MILE OF ITS BORDER. CNT A LL T RANSIT N EIGHBORHOOD C HARACTERISTIC ( TRANSIT ) T RANSIT A CCESS S HED T OTAL AREA THAT TRANSIT RIDERS FROM THE BLOCK GROUP CAN ACCESS IN 30 MINUTES WITH 1 OR NO TRANSFERS FOR ALL THE TRANSIT STATIONS WITHIN A ¼ MILE OF THE BLOCK GROUP CNT A LL T RANSIT N EIGHBORHOOD C HARACTERISTIC ( TRANSIT ) TAS J OBS T HE TOTAL NUMBER OF JOBS IN THE TAS AREA CNT A LL T RANSIT N EIGHBORHOOD C HARACTERISTIC ( TRANSIT )

Costs Auto ownership and use costs: Derived from research conducted by HUD and DOT using the Consumer Expenditure Survey (CES) from the US Bureau of Labor Statistics. The research is based on the waves of the CES, and costs are estimated for autos up to ten years old. Because expenditures are represented in inflation-adjusted 2010 dollars using the Consumer Price Index for all Urban The factor used is derived from the CES; the average expenditure in 2010 is $2,588 and in 2013 it is $3,271, thus the factor applied is Expenses are then segmented by five ranges of household income ($0-$20,000; $20,000-$40,000; $40,000-$60,000; $60,000-$100,000; and, $100,000 and above)

Costs Transit use costs: From The 2013 National Transit Database (NTD) – directly operated and purchased transportation revenue were used. The transit revenue, as reported by each of the transit agencies, to related geographies where GTFS data were collected; revenue was allocated to the counties served based on the percentage of each transit agency’s bus and rail stations weighted by the number of trips provided within each county served. Each county’s estimated transit revenue was assigned to block groups on the basis of molded percent transit journey to work and the total households.

Household Transportation Cost Model

THANK YOU Peter Haas | | Facebook:CenterforNeighborhoodTechnologyFacebook:CenterforNeighborhoodTechnology